Title: Why Might PET be a Good Surrogate Marker in AD?
1Presented by William Jagust, M.D. at
the November 18, 2002 Peripheral and Central
Nervous System Advisory Committee meeting
2Why is PET a Good Potential Surrogate Marker in
AD?
- PET is a good assay for tissue biochemistry and
physiology that is related to disease process - PET is highly related to cognitive function and
predictive of cognitive decline - PET is sensitive, reliable, and valid as a marker
for AD pathology - PET provides statistically powerful measures of
decline
3PET Tracers Relevant to AD
- Radioligands that bind to acetylcholinesterase
- Reflect cholinergic function
- Radioligands that bind to amyloid
- Reflect unknown characteristics of amyloid
pathology - Fluorodeoxyglucose
- Synaptic activity
4Cholinergic Ligands in AD
Kuhl et al, Neurology 1999
5Markers of Alzheimers Disease Temporal-Parietal
Hypometabolism
C
AD1
AD2
6Posterior Cingulate Hypometabolism
7FDG PET and AD Diagnosis
- Hoffman et al., J Nucl Med 2000 (N22)
- Pathologically confirmed AD
- sensitivity 90, specificity 65
- Silverman et al., JAMA 2001
- Progressive dementia (N146)
- sensitivity 93, specificity 76
- Pathologically confirmed AD (N138)
- sensitivity 94, specificity 73
8FDG-PET Predicts Decline
- Baseline PET predicts MMSE change in AD patients
(Jagust et al., 1996) - Baseline PET predicts decline from
memory-impairment to dementia (Minoshima et al,
1997) - Baseline PET predicts memory decline in
nondemented ApoE4 carriers (Small et al, 2000) - Baseline PET predicts decline from normal to MCI
(DeLeon et al., 2001)
9PET Metabolic Rates Predict Rate of Decline in AD
MMSE Change
r0.58
Post Temp Ctx rCMRglc (mg/100g/min)
10Regional Glucose Metabolism is Correlated with
Cognitive Function
Delayed Recall
Delayed Recall
left temporal lobe
left hippocampus
112-year Glucose Metabolic Decline in Asymptomatic
ApoE 4 Heterozygotes
P lt 0.001
Reiman et al., PNAS 2001
12(No Transcript)
13Asymptomatic Subjects Per Treatment Group to
Detect a Drug Effect in Two YearsPost Cingulate
rCMRglc, 80 Power, p 0.01, two tailed
Number of Subjects
Reiman et al, PNAS 2001
Estimated Drug Treatment Effect
14AD Patients per Treatment Group to Detect a Drug
Effect in 1 yearFrontal rCMRglc, 80 Power, p
0.01, two tailed
Number of Subjects
Estimated Drug Treatment Effect
Reiman et al
15FDG-PET as a Surrogate PositivesLink between
PET and clinical decline
- PET predicts clinical decline
- Biologically plausible on the pathway for AD
- sensitive and specific for pathology
- related to pathology, synaptic function
- correlated with cognition
- Statistically powerful
16FDG-PET as a Surrogate NegativesLink Between
PET and Disease Modification
- Problem
- Inability to distinguish symptomatic from
disease-modifying drug effects - Solutions
- Randomized start or withdrawal design
- Use PET tracers that reflect basic biology of AD
amyloid imaging agents - Assess PET in symptomatic drug treatments
17Technical Considerations
- Standardization of multisite acquisition
- state
- input function
- attenuation correction
- resolution
- Standardization of data analysis
- quantitation metabolic rates, ratios
- regions-of-interest, atlas and voxel-based
18Summary
- FDG PET shows potential as a surrogate but
clearly has not been confirmed - Very sensitive in detecting decline, and
statistically powerful - Strong links with clinical symptoms and
pathology, questions on disease modification - Incorporation of PET in clinical trials will help
address questions